To characterize the molecular basis for the hemostatic defects of dengue infections, a study was conducted in Bangkok, Thailand. Febrile children (n = 68) hospitalized with suspected dengue were enrolled before their clinical syndromes were classified as either dengue fever (DF) or dengue hemorrhagic fever (DHF). Hospital course and outcome were recorded; blood was obtained during the febrile illness (S1), after defervescence (S2), and 1 month after onset of disease (S4). Patients were classified as DF (n = 21) and DHF grades 1, 2, and 3; (DHF1, n = 8; DHF2, n = 30; and DHF3, n = 9). All had marked thrombocytopenia. Bleeding scores were assigned on the basis of bleeding site. Although there was no correlation between bleeding scores and pleural effusion index (a measure of vascular leakage) or bleeding scores and platelet counts, there was a correlation between pleural effusion index and platelet counts. Bleeding scores did not correlate with hemostatic data. Activated partial thromboplastin time was prolonged, with trends toward decreased fibrinogen and increased levels of prothrombin fragment F1.2 in the acute-phase samples. However, no factor level was dramatically decreased. We conclude that most patients with DF or DHF, even without overt hemorrhage, have consumptive coagulopathy. Nevertheless, hemorrhage in dengue without circulatory collapse is most likely due to activation of platelets rather than coagulopathy, which is well compensated. Our data suggest that vascular alteration may be the principal factor involved in the association of thrombocytopenia and hemorrhage with disease severity.